From Complex to Practical: Modeling Personality Typologies for use in Applied Research

• In Event: 3-005 - R-E-S-P-E-C-T Individual Differences: A Methods Approach for Examining Culture and Gender

Integrative Statement

Personality is an individual differences variable that can predict many aspects of human behavior (Digman, 1990; Goldberg, 1990). However, the concept we call personality is in fact a complex latent construct represented by a combination of interconnecting qualities and characteristics (Goldberg, 1990; McRae & Costa, 1987). While this is meaningful in research, from an application standpoint the complexity of personality can be limiting. One potential solution to the complexity of modeling personality is the use of latent profile analysis (LPA; Merz & Roesch, 2011; Ferguson & Hull, 2018). LPA is a person-centered approach to latent variable modeling in the same class of statistical approaches as latent class analysis, mixture modeling, etc. (Bergman, Magnusson, & El-Khouri, 2003; Collins & Lanza, 2010). With LPA, individual scores on personality measures are reduced into latent profiles that represent personality typologies (Bergman, et al., 2003; Ferguson & Hull, 2018; Robins, John, & Caspi, 1998).
The present study explores the use of LPA to model personality typologies and use these typologies to understand behavior, specifically science interest. The specific research questions guiding this study are: How many meaningful profiles of personality exist in a sample of high school students from a large public school district in the southwestern United States? How do these personality profiles relate to measures of science interest for these students?
The total sample size for the study was n = 374. Participant mean age in the present study was 17.69, with a standard deviation of 1.21 years. To assess personality, items from the International Personality Item Pool (IPIP) were selected that generally follow the format of the NEO-PI personality measure (Goldberg, et al., 2006; McCrae & Costa, 1987; Costa & McCrae, 1992). Science interest was assessed using a researcher-developed survey.
LPA models were estimated with an increasing number of latent profiles in an iterative model building process, and the model with three profiles was retained (see Table 1). The means and standard deviations of the three latent profiles from Model 3 and the classification probabilities are presented in Table 2. Students in Profile 1 are termed “Excitable”, with relatively high means across all personality subscales and the highest Neuroticism scores of the three profiles. Profile 2 is termed the “Reserved” profile and contains those with relatively low scores on the five personality subscales. And based on the Profile 3 patterns this profile is termed “Well-Adjusted”, with high Extraversion, Openness, Agreeableness, and Conscientiousness, but lower Neuroticism.
Finally, the relationship between the three personality typologies and science interest was evaluated. Personality typology was found to be a statistically significant predictor of science interest, particularly between students with high science interest and those with low science interest. Students towards the middle of the interest distribution were not clearly differentiated based on personality typology. This study therefore supports the use of LPA as a possible approach to modeling personality data that can then be described in practical terms and used for further prediction and understanding of behavior.

Author

• Sarah L Ferguson, Rowan University
  Presenting Author